The quality of automated news is competitive with human journalists for routine tasks, a new study states. Mario Haim, of LMU Munich, and Andreas Graefe, of Macromedia University, Germany, tested people’s perceptions of both automated and human-written news, and the influence of people’s prior expectations regarding the source.
The authors conducted an online survey in Germany with two experimental studies, both with similar opening questions. In the first experiment, the participants read and rated an article without knowing whether it was written by a journalist or an algorithm. In the second experiment, participants compared and rated an article written by a human to a computer-generated one.
Participants had higher expectations for human-written news regarding the’ journalistic quality and readability. Concerning credibility, participants in both experiments excepted no differences or that automated news would be even more credible, the researchers found. Readers’ expectations of quality were rarely met, the researchers state.
When participants saw only one article, differences in perception of automated and human-written articles were small. When shown two articles at once, participants preferred human-written news for readability. Interestingly, they preferred automated news for credibility. Future studies should track how the quality of automated and human-written news will evolve over time, the authors conclude.
The article “Automated News” was published in Digital Journalism and is available online (free abstract) and via ResearchGate (accepted manuscript version available for download).
Picture: Untitled by Artem Sapegin, license CC0 1.0